Pham Ha Hai, Pham Thi Minh Lanh, Tang Van Lam, Pham Van Ngoc, V.V.

Volshanik

УДК DOI: 10.22227/1997-0935.2017.11.p1-p2

SIMULATION OF NEGATIVE PRESSURE WAVE

PROPAGATION IN WATER PIPE NETWORK
Pham Ha Hai1, Pham Thi Minh Lanh2, Tang Van Lam,
Pham Van Ngoc, V.V. Volshanik
Moscow State University of Civil Engineering (National Research University) (MGSU),
26 Yaroslavskoe shosse, Moscow, 129337, Russian Federation
1
Ho Chi Minh City University of Architecture, 196 Pasteur str., Ward 6, district 3,
Ho Chi Minh City, 70000, Vietnam;
2
Ho Chi Minh City University of Technology, 268 Ly Thuong Kiet str., Ward 14, district 10,
Ho Chi Minh City, 70000, Vietnam;

Subject: factors such as pipe wall roughness, mechanical properties of pipe materials, physical properties of water affect the
pressure surge in the water supply pipes. These factors make it difficult to analyze the transient problem of pressure evolu-
tion using simple programming language, especially in the studies that consider only the magnitude of the positive pressure
surge with the negative pressure phase being neglected.
Research objectives: determine the magnitude of the negative pressure in the pipes on the experimental model. The propa-
gation distance of the negative pressure wave will be simulated by the valve closure scenarios with the help of the HAMMER
software and it is compared with an experimental model to verify the quality the results.
Materials and methods: academic version of the Bentley HAMMER software is used to simulate the pressure surge wave
propagation due to closure of the valve in water supply pipe network. The method of characteristics is used to solve the
governing equations of transient process of pressure change in the pipeline. This method is implemented in the HAMMER
software to calculate the pressure surge value in the pipes.
Results: the method has been applied for water pipe networks of experimental model, the results show the affected area of
negative pressure wave from valve closure and thereby we assess the largest negative pressure that may appear in water
supply pipes.
Conclusions: the experiment simulates the water pipe network with a consumption node for various valve closure scenarios
to determine possibility of appearance of maximum negative pressure value in the pipes. Determination of these values
in real-life network is relatively costly and time-consuming but nevertheless necessary for identification of the risk of pipe
failure, and therefore, this paper proposes using the simulation model by the HAMMER software. Initial calibration of the
model combined with the software simulation results and with the model of experiment showed the differences in confidence
intervals using a model that created valve closure scenarios.

K e y w o r d s: water hammer, water distribution system, method of characteristics, pipe materials, negative pressure
surge, water supply pipe

Fo r c i tat i o n: Pham Ha Hai, Pham Thi Minh Lanh, Tang Van Lam, Pham Van Ngoc, Volshanik V.V. Simulation of Nega-
tive Pressure Wave Propagation in Water Pipe Network. Vestnik MGSU [Proceedings of Moscow State University of Civil
Engineering]. 2017, vol. 12, issue 11 (110), pp.
Вестник МГСУ Том 12 Выпуск 11 (110)

МОДЕЛИРОВАНИЕ ОТРИЦАТЕЛЬНОГО ДАВЛЕНИЯ ВОДЫ

В СИСТЕМЕ ВОДОПРОВОДНЫХ ТРУБ
Фам Ха Хай1, Фам Тхи Минь Лань2, Танг Ван Лам, Фам Ван Нгок, В.В. Волшаник
Национальный исследовательский Московский государственный
строительный университет (НИУ МГСУ), 129337, г. Москва, Ярославское шоссе, д. 26;
1
Архитектурный университет Хошимина, 70000, Вьетнам,
Хошимин, район 3, Уорд 6, ул. Пастер, д. 196;
2
Технологический университет Хошимина, 70000, Вьетнам,
Хошимин, район 10, Уорд 14, ул. Ли Тхуон Кит, д. 268;

Предмет исследования: такие факторы, как шероховатость стенок труб, механические свойства материалов труб,
физические свойства воды влияют на гидравлический удар в водопроводных трубах. Эти факторы затрудняют анализ
переходного процесса давления на простом языке программирования, особенно в исследованиях, изучающих только
положительный удар и не рассматривающих отрицательное давление.
Цели исследования: определить значение отрицательного давления в трубах на экспериментальной модели. Рас-
стояние распространения волны отрицательного давления моделируется сценариями закрытия клапана с помощью
программы HAMMER и сравнивается с экспериментальной моделью для проверки репрезентативности результатов.

1098 © Pham Ha Hai, Pham Thi Minh Lanh, Tang Van Lam, Pham Van Ngoc, V.V. Volshanik
 Simulation of Negative Pressure Wave Propagation in Water Pipe Network С. 1098–1107

Материалы и методы: академическая версия программы Bentley HAMMER используется для моделирования рас-
пространения волны давления из-за закрытия клапана в трубах водопроводной сети. Применен метод характери-
стической теории для решения уравнений переходного процесса изменения давления в трубопроводе. Этот метод
используется в программном обеспечении HAMMER для расчета давления в трубах.
Результаты: метод применяется для водопроводных сетей экспериментальной модели, результаты показывают об-
ласть волны отрицательного давления от закрытия клапана. Таким образом, производится оценка наибольшего
отрицательного давления на водопроводных трубах.
Выводы: эксперимент имитирует водопроводную сеть с узлом потребления с различными сценариями закрытия
клапана, чтобы определить возможность появления максимального отрицательного значения давления в трубах.
Определение этих значений в реальной сети является относительно дорогостоящим и трудоемким, но необходимым
для выявления риска отказа трубопровода, поэтому в настоящем исследовании предлагается использовать имита-
ционную модель. Первоначальная калибровка модели, объединенная с результатами моделирования с помощью
программного обеспечения и с моделью эксперимента, показала различия в доверительных интервалах с использо-
ванием модели, которая создавала сценарии закрытия клапана.

К л юч е в ы е с л о в а : гидравлический удар, система распределения воды, метод характеристик, материал трубы,

отрицательный скачок давления, водопроводная труба

Д л я ц и т и р о в а н и я: Фам Ха Хай, Фам Тхи Минь Лань, Танг Ван Лам, Фам Ван Нгок, Волшаник В.В. Simulation
of Negative Pressure Wave Propagation in Water Pipe Network // Вестник МГСУ. 2017. Т. 12. Вып. 11 (110). С.

INTRODUCTION Table 1 shows some research on water hammer

During the operation of water distribution systems,
the water hammer on the pipe usually occurs due to the
pump’s sudden shut off or closure/opening of the valve
on the pipes [1]. In the water supply networks, the num-
ber of valves on the pipe is much more than the number
of pump stations, thus managing the water hammer and
pressure wave propagation due to the valve closure in
the pipes is relatively complex [2]. In addition, the wa-
ter hammer due to pump power-off often has a high in-
tensity and should be very tightly controlled [3], but the
water hammer due to the valve closure is usually small-
er and sometimes ignored during the pipe operation.
However, the effect of water hammer due to the valve
closure is wider because the pressure wave propagates
all over the pipe network, while the pump’s shut off in-
terrupts only the first area of the network [4]. Therefore,
in this study we will focus on analyzing the valve clo-
sure scenarios on the water supply pipe network.
When the water supply pipe is broken under-
ground, it causes a leakage in the water supply network
systems [5] which not only results in economic loss [6]
but also increases the risk of water contamination in the
pipe during the period of negative pressure surge [7],
which is unsafe for water users. Therefore, simulat-
ing the impact of negative pressure wave propagation,
which is a driving force for contamination of water en-
vironment in the pipes [8] is necessary.
LITERATURE REVIEW
Water hammer is the pressure variation in pipes
when the fluid motion is unsteady under the action of
inertia force [9]. The methods of pressure surge cal-
culation, both experimental and theoretical, have been
gradually improved over time, especially with the de-
velopment of computer technology, which has led to
more diverse results in this research field.
performed from 1977 to 2015. The mathematical mod-
els are used to simulate the water hammer in simple
pipes by the difference method [16], the method of
characteristics or the method of algebra. Wylie used
the method of characteristics and the graphical meth-
od to determine pressure surge values in simple pipes,
some examples of water hammer in the pipelines which
are of the same or different diameter, in branch pipes
or the uniform pipes are programmed in Fortran lan-
guage [10]. This paper concentrates on the methods that
determine water hammer in the water distribution sys-
tems computationally.
The dynamic effect of pipe-wall viscoelasticity
in hydraulic transients was investigated by D. Covas.
The valve losses, the physical properties of water and
the characteristics of the pipe material are taken into
account, the magnitude of pressure surge values are
determined and they are calibrated and verified by the
experimental model and the mathematical model [11]. Вестник МГСУ Том 12 Выпуск 11 (110)
The material used for the experiment was Polyethyl-
ene but when setting up the experimental system, the
mechanical properties of the material influenced the
pressure surge to a lesser extent than the position of the
pipe in the network, the history of pressure transient
and prior deformation of the pipe. In the study [12] the
authors found that the position of the pipe is more im-
portant than the mechanical properties of the material
but the results have not been verified on the actual wa-
ter distribution systems and the specific assessment of
the area of the pressure surge wave propagation in the
pipe network system was not carried out. This study
provided a more comprehensive assessment of the po-
tential for water hammer simulation in the water supply
networks.
M.A. Bouaziz evaluated the effect of pres-
sure transient on the wall of the valve when closing
the valve on “Y” type water pipeline networks with-
out pumps and with a pump. The results show that

1099
 Pham Ha Hai, Pham Thi Minh Lanh, Tang Van Lam, Pham Van Ngoc, V.V. Volshanik

Table 1. Literature review of the methods of water hammer simulations

Year Authors Research Title Methods Tools

Method of
Wylie characteristics;
1977 Fluid Transients [10] Fortran
et al. graphical
method
Experimental
The dynamic effect of pipe-wall viscoelasticity in Experimental;
Dídia Covas model and
2004 hydraulic transients. Part I — experimental analysis model of unsteady
et al. mathematical
and creep characterization [11] flow
equations
The dynamic effect of pipe-wall viscoelasticity in Calibration model
Dídia Covas Method of
2005 hydraulic transients. Part II — experimental analysis and comparing the
et al. characteristics
and creep characterization [12] results.
Experimental
Roman Hydraulic transients analysis in pipe networks by the Method of model and
2006
Wichowski method of characteristics (MOC) [13] characteristics mathematical
model
Predicting risk of water quality failures in distribution
networks under uncertainties using fault-tree
analysis [14] bifurcations and change of conducts
characteristics. The numerical algorithm sustains
the maximum pressure values and therefore, FEM
ABAQUS simulation of gray cast iron pipes with a
superficial defect gives the maximum stresses in the
M.A. Bouaziz different network pipes. The severity of a corrosion Method of Fortran and
2014
et al. crater defect was estimated by calculating the safety characteristics ABAQUS software
factor for the stress distribution at the tip of defects.
It allows the acquisition of the applied notch intensity
factor. To study the effects of the geometry defects,
semi-elliptical defects are deemed to exist up to half the
thickness of the pipe wall. To obtain the value of the
safety factor, the results were fed into the assessment
procedure for the structural integrity (SINTAP
Accuracy and security analysis of transient flows in Method of Fortran and
2015 M. Dallali et al.
relatively long pipelines [15] characteristics ABAQUS software

in presence of the pump the positive pressure value by using a program to assess the risk of the pipe failure
increases by 30 % to 40 % and potentially damages under the influence of maximum pressure. The study
Вестник МГСУ Том 12 Выпуск 11 (110)

the pipe structure [14] bifurcations and change of
conducts characteristics. The numerical algorithm
sustains the maximum pressure values and therefore,
FEM ABAQUS simulation of gray cast iron pipes with
a superficial defect gives the maximum stresses in the
different network pipes. The severity of a corrosion
crater defect was estimated by calculating the safety
factor for the stress distribution at the tip of defects.
It allows the acquisition of the applied notch intensity
factor. To study the effects of the geometry defects,
semi-elliptical defects are deemed to exist up to half
the thickness of the pipe wall. To obtain the value of
the safety factor, the results were fed into the assess-
ment procedure for the structural integrity (SINTAP.
The result of M.A. Bouaziz is continued by M. Dallali
et al [15]. In this research, the author determines the
propagation of pressure surge wave in a single pipeline
which connects reservoir and a valve. Along the pipe-
line, pressure head due to closure of the valve is found
used the method of characteristics to calculate pressure
surge but the simulation is valid only for a long pipe-
line because it does not account for the local losses due
to loops or branching with other pipes.
The graphical method used to calculate water ham-
mer provides the complete results. Using this method
it is easier to observe the pressure transient process in
the pipe but it is only suitable for long pipes and non-
branched pipelines because of the large volume of com-
putations and complexity of the boundary. One of the
methods that determines the pressure surge in pipes is
the method of characteristics and it has been used in re-
cent times. Differential equations can be solved by using
finite element mesh of the pipe in which a water ham-
mer occurs, however, the accuracy of results depends on
the number of finite elements used. In this study propose
how to improve this disadvantage by using the Bentley
software HAMMER to simulate the negative pressure
surge generated during the valve closure.

1100
 Simulation of Negative Pressure Wave Propagation in Water Pipe Network С. 1098–1107

MATERIALS AND METHODS where V — flow velocity, m/s; H — the piezometric

head, m; f — Darcy—Weisbach friction factor; g — ac-
Water hammer was simulated by the method of celeration due to gravity, m/s2; t — the time of valve
characteristic, however, the negative pressure transient closure, s; x — length of pipe, m; a — wave speed, m/s,
in water distribution systems was not determined. The
direction of study is approached by simulation of= water a0 k
a = with a0 , (3)
hammer in water pipe network with many loops and K Dt ρ
1+
pipes have different diameters. In this research creates E d0
the valve closure scenarios on the HAMMER model to
evaluate the propagation distance of the wave and de- where E — elastic modulus of pipe wall, MPa; K —
termine negative pressure value due to closure of the elastic modulus of water, MPa; Dt — inner diame-
valve and the maximum negative pressure value that ter, mm; d0 — thickness of pipe wall, mm; a0 — the
may appear in the pipes. acoustic velocity, m/s; ρ — mass density, kg/m3.
The contents of research proposed and the im- The equations (2), (3) are solved by using the finite
plementation steps are shown in Fig. 1. Set up of the element mesh of Wylie and Streeter. The water ham-
experimental model enables us to observe the negative mer appears during the valve closure. Pressure wave
pressure values generated by closing the valve in the propagation in the pipe was divided into the meshes.
model. We use these values to calibrate the HAMMER Using the method of characteristics what is possible to
model setup in the HAMMER software from which the determine the magnitude of piezometric head and flow
affected area and the magnitude of the negative pres- velocity.
sure value in the system are evaluated through the valve 2. The scenarios of valve closure in the HAM-
closing scenarios. The results of research show the like- MER model
lihood of occurrence of negative pressure values in the We used the theory of water hammer due to the
water distribution system and at the same time suggest valve closure in the water distribution system and
appropriate pipeline operation conditions to protect the Bentley’s HAMMER software (for academic edition
pipeline from the occurrence of this pressure. only) [18]. An experimental model was simulated using
1. Governing equations for the water hammer the software as shown in Figure 2. Properties of pipes in
in pipes the model are shown in Table 2. Hazen—Williams fac-
Assuming that the water is a compressible fluid tors are set up according to hydraulic theory presented
and material of the pipe in the network is elastic, pres- in Water distribution handbook [2], and these factors
sure wave velocities are determined by formula (1). will change after calibration of the model.
Considering the case of unsteady flow [17], the flow of The total length of the pipe network is 78.76 m,
water is often modeled using governing equations (2) including the pipes with diameters from 15 mm to
and (3). 140 mm, HDPE and PVC, with 4 loops and 22 branch
Continuity equation is pipes over the area of 70 m2. The underground reser-
voir supplies water to the network via pump and wa-
a 2 ∂V ∂H ∂H
+V + = 0. (1) ter is collected by canals systems to come back to the
g ∂x ∂x ∂t reservoir. Equipment used in the experimental model
includes: Вестник МГСУ Том 12 Выпуск 11 (110)
Momentum equations are • two flow meters measure input and output flow in
the model, their diameter is 30 mm;
∂H fV V ∂V • 6 pressure sensors including 4 positive pressure
g + + = 0, (2)
∂x 2 Dt ∂t sensors and 2 negative pressure sensors. Two pressure
sensors are set up at the first pipe and at the end of the

Figure 1. Contents of research

1101
 Pham Ha Hai, Pham Thi Minh Lanh, Tang Van Lam, Pham Van Ngoc, V.V. Volshanik

Table 2. Properties of pipes in the HAMMER model

Length, Diameter, Hazen—Williams Length, Diameter, Hazen—Williams

Pipe Material Pipe Material
m mm factors m mm factors
1 2.64 131.80 PVC 110 P-34 0.60 20.40 HDPE 125
2 0.36 80.00 PVC 110 P-35 1.80 27.20 HDPE 125
P-1 0.40 65.00 PVC 110 P-36 1.50 27.20 HDPE 125
P-2 0.10 131.80 PVC 110 P-37 0.05 34.00 HDPE 125
P-3 4.02 131.80 PVC 110 P-38 1.20 34.00 HDPE 125
P-4 0.20 131.80 PVC 110 P-39 0.15 34.00 HDPE 125
P-5 0.70 84.20 PVC 110 P-40 0.40 15.40 HDPE 125
P-6 0.10 84.20 PVC 110 P-41 1.10 15.40 HDPE 125
P-7 0.10 84.20 HDPE 125 P-42 0.70 34.00 HDPE 125
P-8 0.30 15.40 PVC 110 P-43 0.14 27.20 HDPE 125
P-9 0.36 84.20 PVC 110 P-44 0.16 27.20 HDPE 125
P-10 1.00 84.20 PVC 110 P-45 1.00 34.00 HDPE 125
P-11 0.74 84.20 HDPE 125 P-46 1.00 34.00 HDPE 125
P-12 0.20 70.00 HDPE 125 P-47 0.20 34.00 HDPE 125
P-13 0.50 34.00 HDPE 125 P-48 1.16 20.40 HDPE 125
P-14 0.23 34.00 HDPE 125 P-49 1.16 20.40 HDPE 125
P-15 0.40 34.00 HDPE 125 P-50 1.16 20.40 HDPE 125
P-16 1.30 34.00 HDPE 125 P-52 0.05 20.40 HDPE 125
P-17 0.30 34.00 HDPE 125 P-53 0.25 20.40 HDPE 125
P-18 0.30 34.00 HDPE 125 P-54 1.35 20.40 HDPE 125
P-19 1.06 34.00 HDPE 125 P-55 1.30 20.40 HDPE 125
P-20 1.00 34.00 HDPE 125 P-56 1.50 34.00 HDPE 125
P-21 0.30 34.00 HDPE 125 P-57 1.20 34.00 HDPE 125
P-22 1.16 34.00 HDPE 125 P-58 0.30 34.00 HDPE 125
P-23 1.80 34.00 HDPE 125 P-59 2.25 34.00 HDPE 125
P-24 1.80 34.00 HDPE 125 P-60 2.20 34.00 HDPE 125
P-25 1.37 34.00 HDPE 125 P-61 1.50 15.40 HDPE 125
P-26 1.50 34.00 HDPE 125 P-62 1.08 15.40 HDPE 125
P-27 0.30 34.00 HDPE 125 P-63 0.24 20.40 HDPE 125
Вестник МГСУ Том 12 Выпуск 11 (110)

P-28 0.52 34.00 HDPE 125 P-64 0.10 34.00 HDPE 125
P-29 1.18 27.20 HDPE 125 P-65 0.08 34.00 HDPE 125
P-30 1.15 20.40 HDPE 125 P-66 1.00 34.00 HDPE 125
P-31 1.85 27.20 HDPE 125 P-67 1.34 34.00 HDPE 125
P-32 0.35 20.40 HDPE 125 P-68 1.78 34.00 HDPE 125
P-33 0.70 20.40 HDPE 125 P-69 0.62 27.20 HDPE 125
P-70 1.00 15.40 HDPE 125

pipe in the network to measure positive and negative
pressures;
• 1 Grundfos pump with flow rate 60.3 m3/h,
11.4 meters head. It is set up after the reservoir.
• 26 gate valves of different diameters: 8 valves are
set up on the main pipes and 18 valves are located on
the branch pipes.
The HAMMER model is calibrated by the experi-
mental model, the setup of which is shown in Figure 3,
and its results are compared with the information pro-
vided by sensors to verify the difference.
Typical head losses in the pipes are due to rough-
ness of the pipe wall. The experimental model has four
pressure sensors to measure the positive pressure values
during the valve closure. These values serve as an input
for the Darwin Calibrator tool of the WaterGEMs soft-
ware to calibrate the HAMMER model [19]. Based on
the values measured experimentally and after using the

1102
 Simulation of Negative Pressure Wave Propagation in Water Pipe Network С. 1098–1107

Figure 2. Simulation of water pipe network in the HAMMER software

Figure 3. Experimental model of water hammer due to valve closure Вестник МГСУ Том 12 Выпуск 11 (110)

Table 3. The difference between negative pressure values obtained in the experiments and the results of HAMMER model

Valve closure Sensor node Experiment, m HAMMER, m Deviation, m

FCV-2 J-13 –9.98 –9.98 0.00
FCV-3 J-37 –9.35 –9.98 0.63
FCV-4 J-15 –9.35 –9.98 0.63
FCV-5 J-52 3.12 2.15 0.97
FCV-6 J-72 –7.28 –8.05 0.77
FCV-7 J-80 –8.52 –9.98 1.46

optimization tool, the results show that the roughness
value for each pipe in the original HDPE pipe model is
C = 150 and the original PVC pipes have the coefficient
of roughness C = 120 [20]. Table 3 shows the devia-
tion between the resulting pressure at the node, obtained
from the HAMMER model after calibration, and the
pressure in the experimental model. The simulated error
is about 2–5 %, which is an acceptable tolerance range
between the HAMMER model and the actual hydraulic
experimental model.

1103
 Pham Ha Hai, Pham Thi Minh Lanh, Tang Van Lam, Pham Van Ngoc, V.V. Volshanik

RESULTS The time of valve closure is 5 s and the time report

In the HAMMER model the water flowing from
the reservoir through the FCV-2 valve diverges into two
main water pipes by two FCV-3 and FCV-4 valves. The
branch pipelines connecting these pipes have a diam-
eter smaller than or equal to that of the main pipes. The
distribution of flow in each pipe corresponds to color
symbols of Figure 4. Red and pink colors correspond to
the flow rates of 1.68 l/s to 2.8 l/s; blue color is a flow of
about 1.12 l/s to 1.68 l/s, and green color has the lowest
flow from 0.56 l/s to 1.12 l/s.
duration in the HAMMER model is 20 s. The valve clo-
sure on the first main pipe, the second main pipe and the
branch pipes are simulated to determine the propaga-
tion distance of negative pressure wave. Figure 5 shows
some of the possible scenarios in the pipes. The mag-
nitudes of negative pressure values –9.98 m :–5.96 m :
–1.94 m : 2.08 m : 10.12 m are shown in the contour
plot with green:cyan:blue:pink:red colors, respectively.
Note that the first pipe from the pump to FCV-2 valve
always has a red color since the pump maintains the
pressure put on the pipe network.

Figure 4. Color coding for flow and pressure in HAMMER model

Вестник МГСУ Том 12 Выпуск 11 (110)

Figure 5. The propagation of the negative pressure wave during the valve closure: a — the first main pipelines; b — the second
main pipelines; c — the branch pipelines

1104
 Simulation of Negative Pressure Wave Propagation in Water Pipe Network С. 1098–1107

Table 4. The maximum negative pressure that may appear when closing the valve
Object The main pipes The branch pipes
Valve closure FCV-2 FCV-3 FCV-4 FCV-7 FCV-5 F9 F11 F13
D, mm* 34 34 34 34 20.4 34 20.4 15.4
Q, l/s* 2.8 1.76 1.04 1.11 0.18 0.24 0.18 0.1
Pmax, m* –9.98 –9.98 –9.98 –9.98 2.15 –4.87 –7.79 –9.98
Propagation-distance
25 12 10 0.3 0.05 0.05 0.05 1.1
from valve, m

*D, Q — diameter and flow in pipes where valve closure takes place
*Pmax — maximum negative pressure value which may appear after valve closure

Table 4 shows some typical results in the valve
closing scenarios in the main and branch pipes of the
HAMMER model. On the first main pipeline, two clo-
sure scenarios for FCV-2 and FCV-3 valves are pre-
sented. A negative pressure of –9.98 m arises across
the network after the FCV-2 valve with the longest
transmission distance equal to 25 m, which is twice the
effective distance (12 m) due to closure of the FCV-3
valve. The reason for this phenomenon is due to the fact
that FCV-2 valve is on the first pipe of the network,
so that, the flow in the network is zero when the vavle
closure
When the FCV-4 valve is closed on P-19, the neg-
ative pressure is –9.98 m, the longest transmission is
10 m after the valve. As the flow of transport on this
pipeline is smaller, the water pressure transmission
distance is also smaller than in the first main pipeline.
However, the extent of the wave propagation depends
not only on the flow but also on the location of the valve
closure in the network. This result can be seen from that
data obtained when the FCV-7 valve is closed at the end
of the network with a flow through the valve equivalent
to P-19 but Pmax = –9.98 m now appears only 0.3 m after
the valve and drops to –6.89 m.
The FCV-5, F9, F11, and F13 valves are located
on branches with different diameters in the middle of
the network, with the same working pressure as in the
main pipelines but the negative pressure values only ap-
pear in the short pipe after valve closure. The result of
simulations show that the maximum negative pressure
values in Table 5. This is especially true in case of clos-
ing the FCV-5 valve on P-52. Here the pressure drop is
from 10.56 m to 2.15 m and it does not lead to negative
pressure. This phenomenon is due to the smaller flow in
P-52 (0.18 l/s) than in the nearby pipes (0.42 l/s). The
pipe is connected to P-70, the initial working pressure
is greater (11.02 m) and the flow direction goes to the
P-52 pipe, and thus, when the valve is closed, the nega-
tive phase of water hammer is compensated by both the
head and the flow from the rear pipes.
The above analysis shows that the area affected as
well as the magnitude of negative pressure value after
the valve closure depends on the flow in the pipe, the
working position of the pipe in the network, the work-
ing conditions before closing the valve as well as the
connection characteristics of the pipes that have valve
closure. The valve closure scenarios of the HAMMER
model are designed to determine the magnitude of the

Вестник МГСУ Том 12 Выпуск 11 (110)

Table 5. The maximum negative pressure value in the pipes Pn

Valve Valve Valve Valve

Pipe Pn Pipe Pn Pipe Pn Pipe Pn
closure closure closure closure
P-14 –9.98 FCV-2 P-22 –6.94 FCV-4 P-70 –9.98 F9 P-63 –2.13 F13
P-15 –9.98 FCV-2 P-25 –6.31 FCV-4 P-29 –4.87 F10 P-64 –1.15 F13
P-16 –9.98 FCV-2 P-26 –5.24 FCV-4 P-31 –3.72 F10 P-44 –9.37 F13
P-17 –9.98 FCV-2 P-39 –4.11 FCV-4 P-35 –2.31 F10 P-38 –9.98 F14
P-18 –9.98 FCV-2 P-69 –4.11 FCV-4 P-36 –0.65 F10 P-43 –9.98 F14
P-42 –9.98 FCV-3 P-52 –4.88 FCV-5 P-30 –8.00 F11 P-56 –9.98 F14
P-45 –9.98 FCV-3 P-32 –4.88 FCV-5 P-34 –5.39 F11 P-57 –9.98 F14
P-46 –9.98 FCV-3 P-65 –5.10 FCV-6 P-40 –5.39 F11 P-59 –8.81 F15
P-28 –9.98 FCV-3 P-27 –9.54 FCV-7 P-33 –9.98 F12 P-60 –6.77 F15
P-47 –9.98 FCV-3 P-58 –9.98 FCV-7 P-54 –7.02 F12 P-68 –9.98 F15
P-48 –9.98 FCV-3 P-20 –9.52 FCV-7 P-55 –3.48 F12 P-21 –5.10 F15
P-49 –9.98 FCV-3 P-66 –6.95 FCV-7 P-41 –9.98 F13 P-23 –3.56 F15
P-50 –9.98 FCV-3 P-37 –9.98 F9 P-61 –7.85 F13 P-24 –1.87 F15
P-19 –7.63 FCV-4 P-53 –9.98 F9 P-62 –4.15 F13 P-67 –4.74 F15

1105
 Pham Ha Hai, Pham Thi Minh Lanh, Tang Van Lam, Pham Van Ngoc, V.V. Volshanik
maximum negative pressure value that can appear in
each pipe of the water supply network in the model.
15 valve closing scenarios give 15 pressure values
for each of the pipes. Usually the maximum negative
pressure value occurs immediately after the valve clo-
sure, but for some pipes in the middle of the network
this value occurs when closing the FCV-2 valve on the
first main pipeline. In actual operation of the network,
the FCV-2 valve closure occurs rarely.
CONCLUSION
The experiment simulates the response of a water
distribution system with a consumption node for differ-
ent valve closure scenarios to determine the magnitude
of the negative pressure value that may occur due to
valve closure. Analysis of the results indicates that the
negative pressure value may or may not appear at the
water hammer phase in front of the valve, and this de-
pends on the magnitude of the working pressure before
closing the valve and, more importantly, the connection
characteristics of the valve. After the valve closure the
probability of appearance of the negative pressure value
is higher than before the valve closure. However, the
magnitude of negative pressure value is different.
REFERENCES
1. Jung B.S., Karney B.W., Boulos P.F., Wood D.J. The
need for comprehensive transient analysis of distribution sys-
tems. American Water Works Association. 2007, vol. 99, no. 1,
pp. 112–123.
2. Mays L.W. Water Distribution Systems Handbook.
Arizona, McGraw-Hill, 1999. 560 p.
3. Lanh Pham Thi Minh, Hai Pham Ha, Truong Nguyen
Вестник МГСУ Том 12 Выпуск 11 (110)
Quang, Hong Le Dinh. Evaluate the possibility of cracking
pipe on water supply network under the age of pipe. Journal
of Environmental Science and Engineering. 2016, B 5, vol. 5,
no. 7, pp. 323–327.
4. Rezaei H., Ryan B., Stoianov I. Pipe failure analysis
and impact of dynamic hydraulic conditions in water sup-
ply networks. Procedia Engineering. 2015, vol. 119, no. 1,
pp. 253–262.
6. Shamir U., Howard C.D.D. An analytic approach to
scheduling pipe replacement. American Water Works Associa-
tion. May 1979, vol. 71, no. 5, pp. 248–258.
7. Saint-Martin E., Kleiner Y., Sadiq R. Predicting risk
of water quality failures in distribution networks under uncer-
tainties using fault-tree analysis. Urban Water Journal. 2008,
vol. 27, no. 1, pp. 1–12.
8. Mansour-Rezaei S., Naser G., Sadiq R. Predicting the
potential of contaminant intrusion in water distribution sys-
tems. American Water Works Association. 2014, no. February,
pp. 105–115.
1106
According to (3) formulation, the magnitude of
the negative pressure surge is inversely proportional to
the pipe diameter, but the simulation on the HAMMER
model indicates that this value depends not only on the
diameter but also on the conditions of the pipes opera-
tion in the network. The negative pressure wave after
valve closure in the main pipelines and in the first pipes
in the network always has the highest and the largest
propagation space. The connection pipes between the
main pipelines are smaller and the propagation distance
is not as wide as in the main pipelines.
The HAMMER software is an option for simula-
tion of the water hammer in water distribution system.
After calibration between software simulation and ex-
periment, we showed that the negative pressure values
are different by 0 m to 1 m, depending on reliability
of the HAMMER model to represent the valve closure
scenarios for the each of the pipes. The pipes can be
protected from the water hammer and the negative pres-
sure values can be reduced by increasing the time of
valve closure and decreasing working pressure before
closing the valve. The main pipelines in the water dis-
tribution systems have higher probability of appearance
of negative pressure values than the branch pipelines.
Therefore, these pipes need to be controlled for pressure
and water quality when the valve closure occurs.
9. Kixêlep P.G., Altsul A.D., Lưu Công Đào. Handbook
of Hydraulic Calculations. Hanoi, Construction Publishing,
2010. 542 p.
10. Wylie E.B., Streeter V.L. Fluid Transients. Shep-
herdstown, WV, McGraw-Hill, 1977. 250 p.
11. Covas D., Stoianov I., Ramos H. et al. The dynam-
ic effect of pipe-wall viscoelasticity in hydraulic transients.
Part I — experimental analysis and creep characterization.
Journal of Hydraulic Research. 2004, vol. 42, no. 5, pp. 517–
532.
12. Covas D., Stoianov I., Mano J.F. et al. The dynam-
ic effect of pipe-wall viscoelasticity in hydraulic transients.
Part II — experimental analysis and creep characterization.
Journal of Hydraulic Research. 2005, vol. 43, pp. 56–70.
13. Wichowski R. Hydraulic transients analysis in pipe
networks by the method of characteristics (MOC). Archives
of Hydro-Engineering and Environmental Mechanics. 2006,
vol. 53, no. 3, pp. 267–291.
14. Bouaziz M.A., Guidara M.A., Schmitt C. et al. Water
hammer effects on a gray cast iron water network after adding
pumps. Engineering Failure Analysis. 2014, vol. 44, pp. 1–16.
15. Dallali M., Guidara M.A., Bouaziz M.A., Schmitt C.
et al. Accuracy and security analysis of transient flows in rel-
atively long pipelines. Engineering Failure Analysis. 2015,
vol. 51, pp. 69–82.
16. Záruba J. Water hammer in pipe-line systems. Am-
sterdam, Elsevier, 1993. 450 p.
 Simulation of Negative Pressure Wave Propagation in Water Pipe Network С. 1098–1107

17. Samri H., Bahrar B., Gueraoui K., Akef F. A poly-
nomial expansion of axial velocity profiles to solve transient
laminar flow in elastic pipe. Applied Mathematical Sciences.
2015, vol. 9, no. 9, pp. 447–457.
18. Mays L.W. Hydraulic Design Handbook. Arizona,
MaGraw-Hill, 1999, 512 p.
19. Bentley. Transient Analysis and Modeling in Ham-
mer. Hammer Con., Bentley, Ed. Bentley system, 2016. 450 p.
20. Bentley Systems. Water distribution modeling and
management — WaterGEMS from Bentley. 2012. 312 p.

Received June 10, 2017.

Adopted in revised form on September 1, 2017.
Approved for publication on October 27, 2017.

A b o u t t h e a u t h o r s : Pham Ha Hai — Candidate of Technical Sciences, Lectuer, Ho Chi Minh City Univer-
sity of Architecture, 196 Pasteur str., Ward 6, district 3, Ho Chi Minh City, 70000, Vietnam; vivmgsu_pham@mail.ru;
Pham Thi Minh Lanh — postdraduate student, Faculty of Civil Engineering, University of Technology, 268 Ly
Thuong Kiet str., Ward 14, district 10, Ho Chi Minh City, 70000, Vietnam; phamthiminhlanh@gmail.com;
Tang Van Lam — Postgraduate Student, Department of Technology of Binders and Concretes, Moscow State
University of Civil Engineering (National Research University) (MGSU), 26 Yaroslavskoe shosse, Moscow,
129337, Russian Federation; lamvantang@gmail.com;
Pham Van Ngoc — Postgraduate Student, Department of Hydraulics and Hydraulic Construction, Moscow
State University of Civil Engineering (National Research University) (MGSU), 26 Yaroslavskoe shosse, Moscow,
129337, Russian Federation; pvngoc12@gmail.com;
Volshanik Valeriy Valentinovich — Doctor of Technical Sciences, Professor, Department of Hydraulics and Hy-
draulic Construction, National Research Moscow State University of Civil Engineering (MGSU), 26 Yaroslavskoe
shosse, Moscow, 129337, Russian Federation; tvg1806@gmail.com.

Поступила в редакцию 10 июня 2017 г.

Принята в доработанном виде 1 сентября 2017 г.
Одобрена для публикации 29 октября 2017 г.

О б а в т о р а х : Фам Ха Хай — кандидат технических наук, предподаватель кафедры воды и окружающей

среды, Архитектурный университет Хошимина, 70000, Вьетнам, Хошимин, район 3, Уорд 6, ул. Пастер,
д. 196; vivmgsu_pham@mail.ru;
Фам Тхи Минь Лань — аспирант кафедры гражданского строительства, Технологический университет
Хошимина, 70000, Вьетнам, Хошимин, район 10, Уорд 14, ул. Ли Тхуон Кит, д. 268; phamthiminhlanh@gmail.com;
Танг Ван Лам — аспирант кафедры технологии вяжущих веществ и бетонов, Национальный исследо-
вательский Московский государственный строительный университет (НИУ МГСУ), 129337, г. Москва,
Ярославское шоссе, д. 26; lamvantang@gmail.com;
Фам Ван Гнок — аспирант кафедры гидравлики и гидротехнического строительства, Национальный ис- Вестник МГСУ Том 12 Выпуск 11 (110)
следовательский Московский государственный строительный университет (НИУ МГСУ), 129337, г. Мос­
ква, Ярославское шоссе, д. 26; pvngoc12@gmail.com;
Волшаник Валерий Валентинович — доктор технических наук, профессор кафедры гидравлики и ги-
дротехнического строительства, Национальный исследовательский Московский государственный строи-
тельный университет (НИУ МГСУ), 129337, г. Москва, Ярославское шоссе, д. 26; tvg1806@gmail.com.

1107